1130 that a symmetry-forbidden transformation proceeds in the presence of a transition metal catalyst should not imply the intervention of the forbidden-to-allowed process. Stepwise catalytic processes reasonably explain the nickel-catalyzed cycloaddition of butadienez5 and the rhodium-catalyzed valence isomerization of cubane, l9 for example. The overall chemistry associated with the catalysis of symmetry-forbidden reactions will unquestionably involve a number of distinguishable stepwise processes in addition to the special case addressed here. In describing postulated stepwise reaction paths, particular attention has been directed toward the “nonconcertedness” of the catalytic process. l 9 t z 3 All molecular transformations to distinct intermediates, of course, are concerted processes, and orbital symmetry restraints necessarily intervene. In the catalysis of [2 21 cycloaddition, symmetry restraints similar in kind to those discussed here can appear.z6 The ligand-
+
(25) P. Heimbach and H. Hey, Angew. Chem., Znt. Ed. Engl., 9, 528 (1970). (26) Consider the oxidative-cycloaddition mechanism proposed for both olefin cyclobutanation and cyclobutane ring-opening processes. 19 For the ring-opening path, the oxidative-addition step, to be symmetry allowed, should be considered a [d2a .2.] process (1) (the subscripts d and a on the first term refer to a metal d orbital which is antisymmetric to the symmetry plane bisecting the two reaction participants). The metal thus supplies an electron pair through one of its antisymmetric atomic orbitals. The second step, however, corresponds to a r.2, a2s .2,] process and the metal thus must withdraw from the transforming ligand system an electron pair through one of its symmetric (with respect to the plane of symmetry passing through metallo-ring) atomic orbitals. Orbital symmetry restraints due to the ligand-field effects discussed above could enter here, introducing energy barriers to the second step. The transition metal, however, has a larger number of symmetric atomic orbitals than antisymmetric and thus can be in a variety of ligand fields which split the d orbitals in a way allowing the unhindered introduction of valence electrons into a symmetric atomic orbital. This would particularly be the case for metal systems with fewer d electrons, where a broader number of symmetric atomic orbitals would be available to accommodate the returning electron pair, The reverse, however, is true for reactions proceeding in the opposite direction. The oxidativeaddition step now is described [& ,2, .2,] and the second step
+
+
+
+
+
field restraints encountered by the stepwise cycloaddition process are essentially those that would be associated with the forbidden-to-allowed process proceeding via monodentate coordination.21 The two paths thus exhibit quite similar orbital symmetry patterns in their modes of catalysis. We feel that the forbidden-to-allowed process will play an important role in the catalysis of symmetryforbidden reactions. Catalytic systems are presently known which are best interpreted through forbiddento-allowed concepts;2dfthe striking ease of these catalytic transformations suggests that significant levels of activity are achievable with appropriate metal systems. Indeed, most unusual catalytic systems have been reported in which trace elements (apparently entrained during purification of reagents over stainless steel spinning bands) catalyze the valence isomerization of quadricyclene derivatives to norbornadiene products, and do so with remarkable fa~ility.~’The nature of these elements is unknown. However, the apparent high levels of activity pose some interesting questions regarding their mode of catalysis. The real breadth of the chemistry associated with the forbidden-to-allowed process is essentially unknown, remaining the subject of experimental research. The dynamics of this process would seem to depend on the thermodynamic driving force of the [ 2 21 ligand transformations, the changes in metalto-ligand coordinate bonding, and the attendant energy barriers associated with the ligand fields of the nonreacting ligands.
+
+
(extrusion) [,2, .2.]. The second step, extrusion of a cyclobutane ring, can experience significant orbital symmetry restrictions due to ligand-field effects, particularly in the d-electron-rich metal systems, since the metal must withdraw a n electron pair through an antisymmetric orbital. (27) P. G . Gassman, D. H. Aue, and D. S. Patton, J. Amer. Chem. Soc., 89, 2486 (1967).
Preparation and Spectroscopic Properties of Methylamino Derivatives of Some Difluoro- and Bis ( trifluoromethyl ) phosphorus Compounds R. G . Cavell,* T. L. Charlton, and W. Sim Contribution from the Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada. Received April 6, 1970 Abstract: The preparation and characterization of four new compounds of the type (CF&P(E)N(H)CHs and F2P(E)N(H)CH3(E = 0, S) by aminolysis of the appropriate chlorophosphoruscompound is described. Infrared and nmr spectra of the compounds indicate that no conformational preference is adopted on the latter time scale although the presence of conformational isomers may be indicated by infrared. Some hydrogen bonding is indicated by infrared studies especially for the compound F2P(0)N(H)CH3,which appears to be strongly hydrogen bonded in all states. The nmr spectral behavior indicates the presence of coupling of the N-H proton to the methyl proton. Second-order intensity variations are observed when the N-H chemical shift is close to that of the CH3 group. The nmr spectra of the phosphines X2PN(H)CH3(X = F, CF3) show the same features.
R
ecently much interest has developed in the structural properties of P-N compounds especially in connection with the question of T bonding and re-
stricted rotation about the phosphorus-nitrogen bond in amino derivatives of trivalent phosphorus compounds. 1-6 As part of a continuing study on the prop-
Journal of the American Chemical Society / 93:5 / March 10, 1971
1131 Table I. Synthesis of Methylaminophosphorus Compounds Reaction quantities-EPXzCl EPXr EPX2Cl reAmine NHCH, taken, covered, taken, yield, mmol mmol mmol mmol(%) OPFzN(H)CH3 OPFzN(D)CHa SPF*N(H)CH3 SPFzN(D)CHa PFzN(H)CHa PFzN(D)CHa (CF&P(O)N(H)CH3 (CF3)2P(O)N(D)CH3 (CFs)*P(S)N(H)CH3 (CFs)zP(S)N(D)CH3
5.4 2.95 6.11 3.09 6.05 4.10b 1.01 0.99 1.06 0.95
0.4 0
0.33 0.8 0 0.50 0 0
A trace of PF3 was also obtained. mined as CF3H.
10.3 6.0 11.8 6.07 10.9 4.14 2.03 1.04 2.25 2.0
Analyses,
?
Calcd N
P
4.67(94)
12.17
26.96
4.85 (84)
10.69
23.7
S
CF3C
24.4
Found P
N 12.00
26.28
10.46
22.8
S
CFac
23.1
5.05(96)" 0 . 4 (40) 0.49 (94) 0 . 9 9 (93)
29.9
29.7
PF3 was used in place of PFzC1 for this synthesis only. Unreacted PF3 was recovered.
erties of volatile fluorophosphorus compounds, we now wish to report the results of synthetic and spectroscopic studies of some four-coordinated pentavalent monomethyl aminophosphorus fluorides and their trifluoromethyl analogs and further studies on related phosphines.
Experimental Section Standard vacuum techniques using Pyrex glass apparatus were employed throughout. Stopcocks were lubricated with Apiezon "N" grease. Infrared spectra were obtained with a Perkin-Elmer 421 dual-interchange (4000-300 cm-l) instrument. Low-temperature spectra were obtained with a cell similar to that described by Wagner and H ~ r n i g . Mass ~ spectra were obtained with an AEI-MS-9 double-focusing mass spectrometer, and nmr spectra with Varian A-56/60 or HA-100 instruments. Proton spectra were measured at 60 and 100 MHz relative to TMS. Fluorine spectra were measured at 94.1 or 56.4 MHz relative to CFCl, and phosphorus spectra were measured at 40.5 MHz relative to P40a. (PBr3was used as a reference to obtain low temperature 31P spectra.) Vapor pressures were measured with a glass spiral microtensimeter8 using the null-point technique with both ascending and descending temperature. Materials. Chlorophosphoryl and chlorothiophosphoryl difluorides were prepared by partial fluorination of phosphoryl or thiophosphoryl trichloride with antimony trifluoride, followed by vacuum fractionation. Methylamine-& was prepared by several successive exchanges of monomethylamine with concentrated NaOD in D20 solution and checked by ir, molecular weight (mol wt calcd for CH3ND2, 33.0; found, 33.0), and mass spectroscopy including mass measurement of the parent ion ( m / e calcd for CHDDZN,33.0548; found, 33.0546). Trifluoromethylphosphorus iodides were obtained from the reaction of CFaI with phosphorus.g Chlorobis(trifluoromethyl)phosphine,10 chlorobis(trifluoromethy1)phosphine sulfide,Il and chlorobis(trifluoromethy1)phosphine oxide1* were prepared by literature methods as was (1) N. N. Greenwood, B. H. Robinson, and B. P. Straughan, J . Chem. SOC.A , 230 (1968). (2) A . H. Cowley, M. J. S. Dewar, and W. R. Jackson, J . Amer. Chem. SOC.,90, 4185 (1968). (3) D . Imbery and H. Griebolin, Z . Naturforsch. B, 23, 759 (1968). (4) M. P. Simmonnin, J. H. Basselier, and C. Charrier, Bu11. SOC. Chim. Fr., 3544 (1967). (5) H . Goldwhite and D . Rowsell, Chem. Commun., 713 (1969). (6) A. H . Cowley, M. J. S. Dewar, W. R. Jackson, and W. B. Jennings, J . Amer. Chem. SOC.,92, 1085 (1970). (7) E. L. Wagner and D . F. Hornig, J . Chem. Phys., 18, 296, (1950). (8) R . Cooper and D . R. Stranks "Techniques of Inorganic Chemistry," Vol. 6, H. B. Johanssen and A. Weissberger, Ed., Interscience, New York, N. Y., 1966, p 1. (9) F. W. Bennett, H. J. Emeleus, and R. N. Haszeldine, J . Chem. SOC.,1565 (1953). (10) A. B. Burg and J. F. Nixon, J . Amer. Chem. SOC.,86, 356 (1964). (11) (a) R. C. Dobbie, L. F. Doty, and R. G. Cavell, ibid., 90, 2015 (1968); (b) K. Gosling and A . B. Burg, ibid., 90, 2011 (1968). (12) J. E. Griffiths and A. B. Burg, ibid., 84, 3442 (1962).
Cacell, Charlton, Sim
Deter-
(CF3)2PN(H)CH3.13The N-deuterated analog of the latter was prepared by the identical procedure using CH3ND2in place of monomethylamine. Preparation of Monomethylamino Derivatives. Method A. Methylamine (2 mol equiv) was condensed into a stopcocked side arm of a 1-1. reaction vessel and 1 mol equiv of the appropriate chlorophosphorus difluoride was condensed into the reaction vessel and allowed to expand into the 1-1. volume. The amine was warmed and slowly admitted to the gaseous chlorophosphorus compound at room temperature through the stopcock provided on the side arm. Reaction occurred immediately on contact of the gaseous reagents with the visible formation of a white solid. After allowing 15 min for complete mixing of the reactants, the volatile materials were admitted to the vacuum system and the constituents separated by fractional condensation. The desired products were trapped at -45", except in the case of F2PN(H)CH3,which passed -63" and was trapped at -96". Excess chlorophosphorus compound was usually used because of easier separation of this constituent from the products. Reactant and product quantities are given in Table I. Method B. Methylamine (2 mol equiv) and 1 mol equiv of chlorophosphine were condensed into a sealed tube which was then allowed to warm to room temperature, whereupon visible reaction was immediately observed. The volatile materials were taken into the vacuum system and separated as above. Vapor pressures of the thiophosphoryl compounds were determined. The data are given in Table I1 and the volatility constants in Table 111. Mass spectral results of all new compounds are given in Tables IV and V. Table 11. Vapor Pressure Data .~
Temp, "C P(obsd), mm P(calcd),b mm Temp, "C P(obsd),mm P(calcd),bmm Temp, "C P(obsd), mm P(calcd),b mm
SPF*N(H)CHz 0.0 6.7= 9 . 9 2.3 3.5 4.0 2.3 4.2 3.5 30.9 3 5 . e 39.8 14.2 17.0 21.5 13.9 17.7 21.8
~
15.5a 6.0 5.9 45,7a 28.7 29.1
~
20.8 8.2 8.0 52.0 39.8 39.0
25.3a 10.6 10.3
(CF&P(S)N(H)CHa 19.8 25.7 32.0 36.0 39.2 45.5 51.2 59.5 6 9 . 0 4 . 4 6 . 2 8 . 7 10.8 12.6 16.1 23.1 31.6 48.2 4 . 5 6 . 2 8 . 6 10.6 12.5 16.9 22.1 32.1 48.2
a Observed with temperature descending from maximum value obtained. All others measured with ascending temperature. Pressures calculated from the equation log P (mm) = A - B/T, using the volatility constants given in Table 111.
Method A was used to prepare F*PN(H)CHa (also prepared elsewhere by a different method14), FzP(O)N(H)CHs, FzP(S)N(H)CH3, and their deuterated analogs because method B consistently gave poor yields of the desired monosubstituted product and (13) G. S. Harris, J . Chem. SOC., 512 (1958). (14) C. G. Barlow, R. Jefferson, and J. F. Nixon, ibid., A , 2692(1968).
1 Deriuatiues of DiJiuoro- and Bis(trifluoromethyl)phosphorus Compounds
1132
(b)
la)
Figure 1. Gas-phase (300°K) and solid-phase (80°K) infrared spectra of the N-H stretching region of (CF&PN(H)CH3 and F2P(S)N(H)CHa.
aminolysis of the P-Cl group can be straightforwardly applied to the CF3 system by simple combination of the required reagents. In order to obtain good yields of the fluoro compounds, it was necessary to conduct the reaction in the gas phase with the chlorophosphorus compound in excess; otherwise concomitant aminolysis of the P-F groups occurs, which yields undesired products. Although P-Cl reacts preferentially with methylamine it is well known that P-F groups also react readily with amines; 15-17 hence it is necessary to dilute the reagents in order to obtain good yields of the desired products. In one case, synthesis of PF2N(D)CH3, readily available PF, was used in place of PF2C1. An alternate route to the synthesis of N-methyl(difluorophosphory1)amine was provided by the gasphase reaction of monomethylamine with difluorophosphoric anhydride O H
(0PFz)zO significant proportions of disubstituted products. With reaction temperatures as low as -95", method B gave improved yields of the desired products but the reactions were not as clean as those done in the gas phase. Reagent quantities, yields, and characterizations are given in Table I.
II I + 2CH3NH2 +F~P-N-CHI + (CHs)NHa+OzPFs- (2)
An alternative preparation of methylaminophosphoryl difluoride involved reaction of difluorophosphoric acid anhydride, (0PFz)zO (0.99 g, 5.31 mmol), with methylamine (0.34 g, 10.97 mmol) in the gas phase according to method A. Immediate reaction was observed upon mixing of the reagents. The only volatile product obtained was F2P(0)N(H)CH3 (0.44 g, 3.79 mmol, 7 2 z yield). A white solid and a small amount of an involatile colorless liquid remained in the reaction vessel. The compounds (CF&P(0)N(H)CH3, (CF3)zP(S)N(H)CH3,and their deuterated analogs were prepared by method B. Reactant and product quantities are given in Table I.
a reaction which is analogous to the synthesis of F2P(0)NH2 from (OPF2)20and ammonia1* except that the latter was done in solution rather than the gas phase. All of these compounds were clear colorless liquids of relatively low volatility. The thiophosphoryl compounds were the most volatile, as is usually observed. The low volatility of the fluorophosphoryl amines made their handling in the vacuum system slow and inconvenient. B. Nmr Spectral Properties. The nmr spectral properties of these compounds are of interest because of the importance of this technique in conformational The methyl-region spectra (7 % 7.0) of F2P(0)N(H)CH3 and (CF3)2P(0)N(H)CH3 are perfectly symmetric and are satisfactorily interpreted by first-order analysis. The methyl-region spectrum of F2P(S)N(H)CH, is not symmetric, but shows an intensity bias, as illustrated in Figure 1; otherwise, the spectral pattern is similar to the symmetric pattern displayed by the above molecules. The fluoro compounds typically show triplet structure in all of the above cases due to 41FH, whereas the jJFHcoupling was not resolved in any of the CF, compounds studied here or elsewhere. A notable feature of all of the spectra is the presence of a coupling of the amino hydrogen (H') with the methyl hydrogens ( 3JHHf) of the order of 5 cps, even though the N-H' signal is broad and poorly defined. The effect of 35HH! and 3JpH couplings is to split the methyl-group resonance into a doublet of doublets. This interpretation is further confirmed by the observations that (1) the spacings within the methyl doublet of doublets are identical at 60 and 100 MHz, indicating that the spectral pattern arises from coupling interactions and not chemically shifted isomers, and also (2) this pattern is not observed in the N-deuterated aminomethyl compounds; rather the X2P(E)N(D)CH3 analogs show only a doublet for the methyl resonance with the 3 J p H COUpling nearly identical with that found in the N(H)CH3
Results and Discussion A. Preparations. The general synthetic reaction
(15) R. G. Cavell, Can. J . Chem., 46, 613 (1968); 45, 1308, (1967); J . Chem. SOC.,1992 (1964). (16) G. Olah, A. A. Ostwald, and S. Kohn, Justus Liebigs Ann. Chem.,
Table 111. Volatilities of Thiophosphorylamines
Volatility constants A B AH,.,,, cal/mol Extrapolated bp, "C Trouton constant, eu
Table IV.
8.065 2105 9633 132.9 23.7
Mass Measurement of Parent Ions Calcd 114.9999 116,0062 130.9761Q 131 .9834a 99.0050 100.0113 214.9935 215.9998 230.9707" 23 1 ,9769
FzP(O)N(H)CH3 FzP(O)N(D)CHa FzP(S)N(H)CHs FzP(S)N(D)CHa FzPN(H)CH3 FzPN( D)CH 3 (CF&P(O)N(H)CH3 (CF3)zP(O)N(D)CHa (CFa)zP(S)N(H)CH3 (CFa)zP(S)N(D)CH3 a
7.836 2105 9634 151.7 22.7
mle
Found 114,9991 116,0062 130.9759O 131.9830" 99,0050 100.0108 214.9936 215.9990 230.9706" 23 1.9754
Accurate masses calculated and measured for 32Sisotope.
XzP(E)Cl
+ 2CH3NHz +X?P(E)N(H)CH, + CHiNHaCl
where X = CF3, F; E
=
(1)
0, S, involving preferential
625, 88 (1959). (17) A. Muller, H. G. Horn, and 0. Glemser, Z . Naturforsch. E, 21, 1150 (1965). (18) S. Kongpricha and W. Preusse, Inorg. Chem., 6 , 1915 (1967).
Journal of the American Chemical Society / 93:5 / March 10, 1971
1133 Table V. Mass Spectra mle
Intensity"
Assignment*
mle
21 5 214 164 147 146 114 107 96 81 80 78 76 69 67 60 59 51 50 49 48 47 46 33 31
(CFa)zP(O)N(H)CH3 2.1 CzFsPONCH4 1.1 CzFaPONCHa 1.6 CFsPFONCHa 0.8 CFaPNCHz 23.3 CFaPONCH4, CF4PNCH 1.8 FZPONCH3, CF3PN 2.0 CFzPNC 12.3 FPONCH4, FzPNCH 0.7 FzPC 15.2 FPNCHa 3.8 FPNCHz 2.3 PONCHs, FPCN 7.5 PFz, CF3 1.3 FPOH 2.4 PNCH3 1.3 PNCHz 1.0 CFzH, PFH 1.0 PF. C K 1.5 HzPO 0.7 HPO 11.2 PO 2.9 PNH 0.7 PHz 1.5 P
133 132 131 130 112 110 103 102 101 98 80 78 69 63 60 50 47 46 45 32
SPFzN(H)CHa 1.4 1.1 28.8 2.6 2.0 2.0 1.7 1.7 8.9 25.6 1.1 1.7 12.9 1.1 0.9 1.7 0.9 1.4 0.9 1.1
M + 2 M S 1 SPFzNCHi SPFzNCHa SPFNCHi SPFNCHz SPFzHz SPFzH SPFz FzPNCH4 FPNCHi FPNCHz PFz PS PNCHs PF PNHz PNH PN S
233 232 231 164 163 162 133 130 119 114 113 112 100 93 92 83 81 80 78 69 65 64 63 61 60 59 51 50 46 32 31 115 114 112 104 98 96 94 87 86 85 69 67 66 50 47
Intensit ya
Assignmentb
(CFa)zP(S)N(H)CHs 0.6 M+2 0.8 M S 1 7.5 CzFsPSNCH4(M) 1.1 162 2 1 1.3 162 16.7 CF~PSNCHI 0.7 CF4PN 1.4 FzPSNCH8 1.3 CFaPF 0.9 CFaPN, 112 2 1.2 1 CFzPS, 112 16.1 FPSNCHI 0.6 CzF4, CF3P PSNCH4 1 .o PSNCHa 1.6 1.7 FzPN 1.1 FzPC 13.9 FPNCHa 1.6 FPNCHz, HNPS 5.6 CFa, PFs FPNH 0.8 1 .o FPN 9.7 PS 1 .o PNCHi 2.8 PNCHa 1.2 PNCHz 0.9 CFzH, PFH 0.9 CFz, PF 2.8 PNH S 1.3 0.8 P
+ +-
+ +
OPFtN(H)CH1 22.2 OPFzNCH4 23.7 OPFzNCHa 2.6 OPFzNCH 2.1 OPFs 1.3 FzPNCHs 2.8 OPFNCH4 5.4 OPFNCHZ 10.8 OPFzHz 2.1 OPFzH 8.2 OPFz 7.7 PK 4.4 POFH 1.3 OPF 1.0 PF 1.8 PO
Ions such as NCH4, NCH,, and NCH, arising from the fragmentation of the methylamino group are not included in the listing. Intensities are expressed relative to the total ionization defined as ZJintensity) for all ions with mass greater than 30 whose intensity is greater than Positive ion spectra. Masses of sulfur-containing fragments are calculated for the 3 3 isotope. The symbol M 2 z of the base peak. refers to the molecular ion.
compounds. Because D-H coupling constants are about one-seventh the size of H-H coupling constants, the former were not usually resolved, although one case, illustrated in the spectrum of F2P(S)N(D)CH3 (Figure l), did show this splitting owing to a fortuitous coincidence of 4JFHand 3JDHwhich gave an easily resolved overlapping pattern. The spectral parameters are given in Table VI. The intensity bias in the methyl nmr spectrum of F2P(S)N(H)CH3 can be ascribed to the proximity of N H and CH3 resonances, whereupon the ratio J / a indicates that second-order spectral interpretation is required. Second-order spectra calculatedlg from the parameters listed in Table VI give both line positions and intensities in excellent agreement with observed (19) Spectra were calculated by the program plied by Dr. J. s. Martin of this department.
NSPECT 111
kindly sup-
Cavell, Charlton, Sim
spectra. The greater chemical shift differences between N-H and CH3 in the compounds (CF3),P(0)N(H)CH3 and F2P(0)N(H)CH3 (where T N H -4.0-5.0) allows a first-order interpretation. We have obtained similar results to the above for (CF3)2PN(H)CH3,in agreement with a recent reinterpretation of the methylregion nmr spectrum of this compound. Thus the doublet of doublets observed' can be ascribed to amino hydrogen and phosphorus coupling with the methyl protons. The constancy of the four-line CHI pattern at both 60 and 100 MHz and the collapse of the fourline pattern to a simple doublet in the N-deuterated derivative supports this interpretation6 and does not support the proposed presence of rotational isomers. The methyl proton spectrum of F2PN(H)CH3is, in agreement with a previous report,I4 a broad complex multiplet at both 60 and 100 MHz for reasons which are
Derivatives of Difluoro- and Bis(trifluoromethyI)phosphorus Compounds
1134 Table VI, Nuclear Magnetic Resonance Parameters of Methylamino Compounds -Chemical shifts.--dNH) dCH3) dF) (CF3)zPN(H)CHad (CF3)zPN(D)CHa (CF~)ZP(O)N(H)CH~ (CF&P(O)N(D)CHa (CFa)zP(S)N(H)CH3 (CFa)zP(S)N(D)CHa FzPN(H)CH3 (in CClIF)' F2PN(H)CH3(in py-CClrF) FzPN(D)CH3 FzP(O)N(H)CHs FzP(O)N(D)CH3 FzP(S)N(H)CHs' F*P(S)N(D)CH3
7.7 4.5 -7 ? ?
?
7.16 7.11 7.08 7.09 7.08 6.99 7.4 7.9 7.2 7.23 7.22 7.07 7.06
64.5 60.8 72.2 72.2 72.3 72.4 70.2 69.5 70.5 80.5 81.5 52.4 52.2
Coupling constantsbpc JPF
JPH
82.4 82.0 111 112 106 106 1197 1177 1190 991 986 1083 1083
JPH
I
11.6 NR 11.4 11.5 NR 11.5 13.2 NR 13.3 Complex 7.05 NR 13.7 20 13.8 14.3 20.7 14.2
JFH
NR 0.6 0.6 0.5 NR 0.6 2.0 2.4 NR 1.2 NR 1 .o
-1
.o
JFH,
JHH,
NR
5.2
NR
5.6
NR
4.6
10.35 11 .o
NR 5.8
1.5
5.6
5.2
5.8
5 Hydrogen chemical shifts are measured relative to tetramethylsilane T 10.0. Fluorine chemical shifts are given relative to CC1,F. b Values NR = not resolved. In agreement with ref 1 and 6. e In agreement with ref 16. f 31P chemical in cps. H' refers to the NH proton. shift = 42.6 ppm 6s. P406.
identical with those outlined above. In this case further evidence is provided by the pronounced simplication of the 60-MHz proton spectrum of F2PN(H)CH3 to a doublet of doublets of triplets upon addition of pyridine to the sample solution. The 19Fspectrum of this solution shows a doublet of doublets of quartets as expected for the above structure in which all possible spin-'/,, couplings are resolved into a first-order spectrum. It seems reasonable to suggest that interaction of pyridine serves to shift the amino proton to lower field, removing the accidental coincidence of chemical shifts which gave rise to the complex multiplet pattern. The coupling of the amino proton to the methyl group is observed as in all other compounds in this system. The spectra of all of the X2P(E)N(H)CH3 compounds with the exception of (CF3),P(S)N(H)CH3 are essentially unchanged with temperature down to about -90", which was the lowest temperature available to us. The methyl-region spectrum of (CF,),P(S)N(H)CH3 did, however, show some temperature variations. At normal probe temperatures (35 ") a doublet (J G 13.4 cps) of relatively broad and nonsymmetric lines is observed. Cooling the sample to about -20" results in the resolution of two components from each broad line: one apparently about twice as intense as the other member. Further cooling improves the separation of the two components until at -40" the spectrum is similar in appearance to that of (CF,),PN(H)CH3 except that the weaker components of the secondorder A, part of the A3BX spectrum appear on the highfield rather than low-field side. Further cooling has little effect on the spectrum except that at -80" a broad signal which can be assigned to the N-H resonance emerges from the low-field side of the methyl multiplet and so accounts for the direction of the intensity bias. The spectrum is best interpreted in a way consistent with that of all of the other X2P(E)N(H)CH3 compounds with the only differences being due to the position of the N-H chemical shift. The temperature behavior of the CH3 spectrum of (CF&P(S)N(H)CH3 is best interpreted as the result of changes of the NH' chemical shift with temperature. The N-H' proton, however, couples with the methyl protons as in the other compounds, creating the characteristic A3BX pattern observed in all cases here. The N-deuterated derivative shows only a sharp doublet which is unJournal of the American Chemical Society 1 93:s
changed in shape and position down to -90", confirming the above interpretation. The 100-MHz spectrum at normal probe temperatures (+40°) is similar in appearance to the 60-MHz spectrum obtained at -20" and shows the characteristic intensity-biased doublet of doublets. The resolution of these features at the ordinary probe temperature (40") of the 100-MHz instrument probably occurs because of the increased separation of CH3 and NH' resonances at the higher field. The sense of the intensity bias is the same as that shown by the -20 to -60" spectra obtained at 60 MHz. All of these compounds therefore exhibit A3BX spectra characterized by coupling of phosphorus and the amino proton to the methyl group. They differ only in the position of the N H chemical shift relative to that of the CH3 chemical shift and its relationship to the coupling constants. Although the N-H resonance is usually well separated from that of CH3, occasional accidental proximity must be expected, particularly when the nitrogen contains such strongly modifying substituents as (CF3)*P or F2P groups. It is possible that the geometry at the nitrogen is planar or nearly so in these compounds as it is in crystalline F2PN(CH3)220 and this feature may be responsible for the observed chemical shifts. The nmr spectra, however, provide no evidence of distinct rotational isomers in any of these compounds down to -90". Recent indications6 that rotational isomers can be distinguished in the nmr spectra of (CF&PN(H)CH3, C12PN(CH3)2, and (CF&PN(CH3) at temperatures of the order of - 120" indicate that the spectra observed here from +40" to -90" are averages of all rotational conformations. The barriers to rotation about the P-N bond in the presently investigated series of compounds are probably similar to the values of 8-9 kcal estimated for the above phosphines.6 We are attempting to extend the temperature range of our measurements in order to obtain a clearer picture of rotational barriers in the pentavalent P(0) and P(S) compounds. C. Infrared Spectral Studies. In general, the gasphase infrared spectra of compounds of the type XtP(E)NHR show two rather than one band in the N-H stretching region. The bands are best described in terms of a poor-to-well-resolved shoulder on a peak of (20) E. D. Morris, Jr., and C.E. Nordman, Inorg. Chem., 8, 1673 (1969).
March 10, 1971
1135
d
E
5m
3
3
3
E
vi m
%
Cacell, Charlton, Sim
m
E
WEE
E
8 v
Derivatives of Dijuoro- and Bis(trifluoromethy1)phosphorus Compounds
1136
'
interactions of the N-H group and differences of interaction in conformational isomers. The lack of change of the spectrum of F2P(0)N(H)CH, with change of state suggests that this compound exists primarily in the hydrogen-bonded form in gas, solution, and solid states. In view of the persistence of the association and the low volatility of the compound, it seems reasonable to propose that the hydrogen-bond association is intermolecular; perhaps through the formation of a dimer (1) or perhaps extended-chain poly-
I
I
1
CALC
L
-
I
Figure 2. A comparison of observed methyl region ( T -7.0) of F2P(S)N(H)CH3at 60 and 100 MHz with the calculated spectrum and with the spectrum of the N-deuterated analog. All spectra were obtained on an approximate 15% solution of the compounds in CC1,F and all spectra are shown on the same sweepwidth scale (100 cps) with field increasing to the right.
reasonable intensity. A typical example is illustrated in Figure 2. Gas-phase spectra of (CF3),P(0)N(H)CH3 and F2P(S)N(H)CH3 have a relatively sharp absorption in the N-H region at values generally assigned to an unassociatedZ1N-H group (3400-3500 cm-l) with a shoulder at lower energy. The gas-phase spectra of (CF3)2P(S)N(H)CH3 and FzP(0)N(H)CH3 are different, however ; the former shows only one fairly sharp band at 3435 cm-l and the latter has a relatively broad band centered at 3256 cm-l. Condensing the above compounds to the solid state at approximately 80°K resulted in three cases in a pronounced broadening of the N-H band and a significant shift to lower wave-number values as shown in Table VII. The relatively broad band at 3256 cm-l in the gas-phase spectrum of FzP(0)N(H)CH3 did not, however, significantly change its shape or position upon condensing the compound to the solid phase at 80°K. A 10% solution of F2P(0)N(H)CH, in CC1, showed the same broad band at 3260 cm-l. The shape and position of this band were unchanged upon dilution of the solution to 0.05%. It seems reasonable to postulate that the observed spectral changes are due to two effects, hydrogen-bonding (21) L. J, Bellamy, "Infrared Spectra of Complex Molecules," 2nd ed, Methuen, London, 1958.
Journal of the American Chemical Society
1 93:5
mers. Intramolecular hydrogen bonding would appear to be less likely because of unfavorable steric requirements. The remaining compounds, including (CF,),P(O)N(H)CH3, do not provide evidence in favor of strong hydrogen bonding in the gas phase; rather the molecules appear to exist as equilibrium mixtures of two conformations in the gas phase as indicated by the presence of two N-H frequencies which can be assigned to two conformations of the N(H)CH3 group with respect to the X2P group. Condensation to the solid state, however, gives rise to substantial shifts in the N-H stretching frequency which are compatible with the formation of hydrogen-bonded N-H groups in the solid. The most marked shift of the N-H frequency upon condensation to the solid state is about 300 cm-' in the case of (CF,),P(O)N(H)CH,, The solid-state N-H absorption is then similar in both phosphoryl compounds. Both of the thiophosphoryl compounds exhibited smaller shifts of 135-180 cm-' upon condensation to the solid state, which suggests that hydrogen bonding to the P=O or P=S group is involved in all systems rather than hydrogen bonding to fluorine atoms. Finally, the lowest shifts are those shown by the phosphines, 55 cm-l from the strongest gas band to the solid-state band in the case of (CF&PN(H)CH, and 42 cm-l in the case of F2PN(H)CH3. The sharpness of the bands in the solid-state spectra of the phosphines suggests that the molecules have been stabilized in one conformation at this temperature, but it is not possible to decide from the available evidence which conformation is preferred. The broadness of the bands in the other compounds, particularly the P=O compounds, may arise from the existence of more than one conformation or from various forms of intermolecular association in the solid state. It is unfortunate that all of our attempts to produce narrow bands in the solid-state spectra of P=O and P=S compounds by means of various annealing and condensation procedures have been unsuccessful. The two bands which are observed in the gas-phase spectra of all XzP(E)N(H)CH3 compounds except (CF3),P(S)N(H)CH3 (which shows only one band) are most probably due to the presence of different rotational isomers in the compounds in which the N-H to the stretching frequency is affected by its substituents on phosphorus. Both frequencies occur
March 10, 1971
1137 in the nonassociated region and may be reasonably assigned to the cis and trans orientations of the N-H group relative to the X2P groupaZ1 The detailed assignments are further complicated by the possibility that the nitrogen may be planar in these molecules;*o thus the bands may arise through the existence of three conformations (Figure 3, I-111), of which two (I and 11) will probably have similar N-H environments and thus similar N-H stretching frequencies, whereas the third pyramidal nitrogen conformation (111) will have a different stretching frequency from the other two. If on the other hand the nitrogen is planar, the two most likely conformations (Figure 3, IV and V) provide cis and trans orientations of the CH3 groups relative to the (CF3)2P group. In the absence of the appropriate dipole moment data a definite interpretation of the intensity patterns cannot be made at this time. If, however, following others,6 we assume a nearly planar nitrogen environment and that steric arguments suggest that the trans-CH, (V) structure is the most abundant and assign the lower frequency to the cis conformation and the higher to the trans,21we are led to the conclusion that the cis form is the least abundant form observed in the spectra of F2PN(H)CH3 and the pentavalent X2P(E)N(H)CH3 compounds except F2P(0)N(H)CH3, which appears to be wholly hydrogen bonded, and (CF3),P(S)N(H)CH3, which shows only one band in the N-H region. In this latter case the observation of only one band suggests that either only one conformation exists (which is unlikely in view of the nmr results which suggest that the N(H)CH3 group is freely rotating on the nmr time scale at 25") or that there is little difference in the N-H stretching frequencies of the cis and trans conformations. In the case of (CF3)2PN(H)CH3the weaker component of the N-H absorptions is found to high energy rather than low energy, in contrast to the behavior of F2PN(H)CH3. As a consequence, either the cis (IV) form must be the more abundant of the two conformations (perhaps through strong intramolecular interactions or substantially different dipole moments in each conformation) or the reverse assignment of the frequencies should be made in this case. The infrared spectrum of the compound (CF3)2PN(H)CH3is also anomalous in the series of molecules studied here in that it shows the largest splitting between the two gas-
Cavell, Charlton, Sim
J.
i'
CH3\
IV
i
V
Figure 3. A representation of the conformations of (CF&PN(HICHI.
phase N-H frequencies. It has been clearly established, however, by overtone and isotopic substitution spectral that both the 3465- and 3430-cm-l bands arise from N-H stretch vibrations. The very pronounced shift of the N-H stretching region to one sharp band in the solid state further suggests that the two bands in the gas phase arise from different conformers, only one of which is favored in the solid. Unfortunately, the available evidence does not permit a decision as to whether the solid structure contains the cis or the trans conformers. It is interesting to note also that the relative intensities of the twc bands in the N H region of the spectrum of (CF3),PN(H>CH3did not change over the temperature range 20-90 O, which suggests that the relative populations are not greatly affected by temperature in this region, in contrast to the pronounced effect observed upon solidifying the compound. Further understanding of the behavior of (CF&PN(H)CH, will only be achieved with additional gas- and solid-state structural data, particularly on the conformation and geometry of the N(H)CH3group in both states. Acknowledgment. We thank the National Research Council and Defence Research Board of Canada for financial support of this research. Also we thank Miss J. K. Schneider for some experimental assistance. We also thank the National Research Council of Canada for a scholarship awarded to T. L. C.
Derivatives of Difluoro- and Bis(trifluoromethyl)phosphorus Compounds
